Artificial Intelligence Platform for speech rehabilitation

The Challenge

Develop a support tool that, through automatic speech processing and graphic animation, complements the work of speech therapists, allowing the patient's rehabilitation treatment to go beyond the consultation time.

The Proposed Solution

Creation of an online platform with Artificial Intelligence, which allows users to pronounce or listen perfectly as pronounced by obtaining immediate feedback, and to speech therapists to monitor and know the real evolution of these patients.

The Result

The platform allows:

• Improves the articulation of phonemes. Obtaining auditory and visual feedback or feedback from the production of phonemes favors a greater and faster generalization of the correct articulation.
• Improves prosody and speech intonation. Prosodic parameters such as the correct placement of the accentuation in the words, the rhythm at the level of syllables and words or the curvature of the intonation from the audio generated by the patient are analyzed.
• Extends the rehabilitation time. Users have the possibility of increasing the treatment time beyond the face-to-face session with the speech therapist, favoring greater autonomy and involvement in the treatment.
• Examples of correct articulation and intonation. The patient has examples of the articulation and correct intonation of multiple stimuli at the auditory and visual level, both in videos and 2D animations.

In addition, the project has addressed important scientific-technological challenges such as the adaptation of speech recognition technology to people with pathologies, the evaluation of pronunciation at the phonetic, articulatory and prosodic levels, establishing a differential diagnosis of the causes of dysfunction or customize the evaluation and individualize the rehabilitation treatment according to the progress of each patient.